search

Illumina / canvas

Canvas - Copy number variant (CNV) calling from DNA sequencing data
Other
121 stars 20 forks source link

System.IO.FileNotFoundException: The specified vcf file () does not exist. #95

Closed ATpoint closed 5 years ago

ATpoint commented 6 years ago

Dear Eric,

might you have a look at the following error message. Somatic-WGS (current release) was run with this code:

$DOTNET $CANVAS Somatic-WGS \
  --bam=${BASENAME}-t_SortedRmdup.bam \
  --somatic-vcf=Strelka2Dir_${BASENAME}/Somatic/results/variants/somatic.snvs.vcf.gz \
  --sample-b-allele-vcf=Strelka2Dir_${BASENAME}/Germline/results/variants/variants.vcf.gz \
  --sample-name=${BASENAME}-t \
  --output=CanvasDir_${BASENAME}/Somatic \
  --reference=${KMER} \
  --genome-folder=${GENOME_FOLDER} \
  --filter-bed=${FILTER13}

Here are both CanvasLog.txt and CanvasError.txt. After checkpoint 5, it complains about Unhandled Exception: System.IO.FileNotFoundException: The specified vcf file () does not exist. Can you help out here?

best wishes, Alexander

eroller commented 5 years ago

looks like some file is getting the empty string. It might actually be the bed file and not a vcf file. Please check all your environment variables, I suspect one of them is not set. Alternatively add set -u to the top of the bash script and it should catch that error.

MatthewPorterGN commented 5 years ago

Hi @eroller,

I am experiencing the same issue, although with Somatic-Enrichment mode. While double-checking my inputs I used filepaths without variable substitution and validated that all input files are present, although my example below shows a variable just to keep the command generic. The inputs seem to get processed fine up until checkpoint 05. The 'missing VCF' mentioned in the error message is produced internally by Canvas, following the name format of:

/data/output/${SAMPLENAME}/TempCNV_${SAMPLENAME}/VFResults${SAMPLENAME}.txt.gz

This VCF exists but it is empty, containing only a header line:

#Chromosome Position Ref Alt CountRef CountAlt

MatthewPorterGN commented 5 years ago

This wound up being an issue with the latest version of Canvas; switching to v1.11.0 as https://github.com/Illumina/canvas/issues/62 suggests fixed the problem. Usage of this version is not mentioned anywhere in the main documentation.

If T/N and Somatic-Enrichment modes are deprecated, why are broken versions of them still included in the final releases? Shouldn't these modes be pulled from the main documentation page, or the documentation updated to point users to the correct version for running these modes, in order to avoid confusion?